Publications
281 results found
Tamaddoni-Nezhad A, Muggleton S, 2011, Stochastic Refinement, 20th International Conference on Inductive Logic Programming (ILP), Publisher: SPRINGER-VERLAG BERLIN, Pages: 222-237, ISSN: 0302-9743
- Author Web Link
- Cite
- Citations: 2
Santos J, Muggleton S, 2010, Subsumer: A prolog θsubsumption engine, Pages: 172-181, ISSN: 1868-8969
State-of-the-art θ-subsumption engines like Django (C) and Resumer2 (Java) are implemented in imperative languages. Since θ-subsumption is inherently a logic problem, in this paper we explore how to efficiently implement it in Prolog. θ-subsumption is an important problem in computational logic and particularly relevant to the Inductive Logic Programming (ILP) community as it is at the core of the hypotheses coverage test which is often the bottleneck of an ILP system. Also, since most of those systems are implemented in Prolog, they can immediately take advantage of a Prolog based θ-subsumption engine. We present a relatively simple (1000 lines in Prolog) but efficient and general θ- subsumption engine, Subsumer. Crucial to Subsumer's performance is the dynamic and recursive decomposition of a clause in sets of independent components. Also important are ideas borrowed from constraint programming that empower Subsumer to efficiently work on clauses with up to several thousand literals and several dozen distinct variables. Using the notoriously challenging Phase Transition dataset we show that, cputime wise, Subsumer clearly outperforms the Django subsumption engine and is competitive with the more sophisticated, state-of-the-art, Resumer2. Furthermore, Subsumer's memory requirements are only a small fraction of those engines and it can handle arbitrary Prolog clauses whereas Django and Resumer2 can only handle Datalog clauses.
Kay E, Lesk VI, Tamaddoni-Nezhad A, et al., 2010, Systems analysis of bacterial glycomes, BIOCHEMICAL SOCIETY TRANSACTIONS, Vol: 38, Pages: 1290-1293, ISSN: 0300-5127
- Author Web Link
- Cite
- Citations: 10
Lodhi H, Muggleton S, Sternberg MJE, 2010, Multi-class Mode of Action Classification of Toxic Compounds Using Logic Based Kernel Methods, MOLECULAR INFORMATICS, Vol: 29, Pages: 655-664, ISSN: 1868-1743
- Author Web Link
- Cite
- Citations: 1
Muggleton S, Paes A, Santos Costa V, et al., 2010, Chess revision: Acquiring the rules of chess variants through FOL theory revision from examples, Pages: 123-130, ISSN: 0302-9743
The game of chess has been a major testbed for research in artificial intelligence, since it requires focus on intelligent reasoning. Particularly, several challenges arise to machine learning systems when inducing a model describing legal moves of the chess, including the collection of the examples, the learning of a model correctly representing the official rules of the game, covering all the branches and restrictions of the correct moves, and the comprehensibility of such a model. Besides, the game of chess has inspired the creation of numerous variants, ranging from faster to more challenging or to regional versions of the game. The question arises if it is possible to take advantage of an initial classifier of chess as a starting point to obtain classifiers for the different variants. We approach this problem as an instance of theory revision from examples. The initial classifier of chess is inspired by a FOL theory approved by a chess expert and the examples are defined as sequences of moves within a game. Starting from a standard revision system, we argue that abduction and negation are also required to best address this problem. Experimental results show the effectiveness of our approach. © 2010 Springer-Verlag Berlin Heidelberg.
Watanabe H, Muggleton S, 2010, Can ILP be applied to large datasets?, Pages: 249-256, ISSN: 0302-9743
There exist large data in science and business. Existing ILP systems cannot be applied effectively for data sets with 10000 data points. In this paper, we consider a technique which can be used to apply for more than 10000 data by simplifying it. Our approach is called Approximative Generalisation and can compress several data points into one example. In case that the original examples are mixture of positive and negative examples, the resulting example is ascribed in probability values representing proportion of positiveness. Our longer term aim is to apply on large Chess endgame database to allow well controlled evaluations of the technique. In this paper we start by choosing a simple game of Noughts and Crosses and we apply mini-max backup algorithm to obtain database of examples. These outcomes are compacted using our approach and empirical results show this has advantage both in accuracy and speed. In further work we hope to apply the approach to large database of both natural and artificial domains. © 2010 Springer-Verlag Berlin Heidelberg.
Lodhi H, Muggleton S, Sternberg MJE, 2010, Multi-Class protein fold recognition using large margin logic based divide and conquer learning, ACM SIGKDD Explorations Newsletter, Vol: 11, Pages: 117-122, ISSN: 1931-0145
<jats:p>Inductive Logic Programming (ILP) systems have been successfully applied to solve complex problems in bioinformatics by viewing them as binary classification tasks. It remains an open question how an accurate solution to a multi-class problem can be obtained by using a logic based learning method. In this paper we present a novel logic based approach to solve complex and challenging multi-class classification problems by focusing on a key task, namely protein fold recognition. Our technique is based on the use of large margin methods in conjunction with the kernels constructed from first order rules induced by an ILP system. The proposed approach learns a multi-class classifier by using a divide and conquer reduction strategy that splits multi-classes into binary groups and solves each individual problem recursively hence generating an underlying decision list structure. The method is applied to assigning protein domains to folds. Experimental evaluation of the method demonstrates the efficacy of the proposed approach to solving multi-class classification problems in bioinformatics.</jats:p>
Lodhi HM, Muggleton SH, 2010, Elements of Computational Systems Biology, ISBN: 9780470180938
Groundbreaking, long-ranging research in this emergent field that enables solutions to complex biological problems. Computational systems biology is an emerging discipline that is evolving quickly due to recent advances in biology such as genome sequencing, high-throughput technologies, and the recent development of sophisticated computational methodologies. Elements of Computational Systems Biology is a comprehensive reference covering the computational frameworks and techniques needed to help research scientists and professionals in computer science, biology, chemistry, pharmaceutical science, and physics solve complex biological problems. Written by leading experts in the field, this practical resource gives detailed descriptions of core subjects, including biological network modeling, analysis, and inference; presents a measured introduction to foundational topics like genomics; and describes state-of-the-art software tools for systems biology. Offers a coordinated integrated systems view of defining and applying computational and mathematical tools and methods to solving problems in systems biology. Chapters provide a multidisciplinary approach and range from analysis, modeling, prediction, reasoning, inference, and exploration of biological systems to the implications of computational systems biology on drug design and medicine. Helps reduce the gap between mathematics and biology by presenting chapters on mathematical models of biological systems. Establishes solutions in computer science, biology, chemistry, and physics by presenting an in-depth description of computational methodologies for systems biology. Elements of Computational Systems Biology is intended for academic/industry researchers and scientists in computer science, biology, mathematics, chemistry, physics, biotechnology, and pharmaceutical science. It is also accessible to undergraduate and graduate students in machine learning, data mining, bioinformatics, computational biology, and systems b
Muggleton S, Santos J, Tamaddoni-Nezhad A, 2010, ProGolem: A System Based on Relative Minimal Generalisation, 19th International Conference on Inductive Logic Programming, Publisher: SPRINGER-VERLAG BERLIN, Pages: 131-148, ISSN: 0302-9743
- Author Web Link
- Cite
- Citations: 12
Muggleton SH, 2010, Knowledge Mining Biological Network Models., Publisher: Springer, Pages: 2-2
Lodhi H, Muggleton S, Sternberg MJE, 2009, Multi-class protein fold recognition using large margin logic based divide and conquer learning, Pages: 22-26
Inductive Logic Programming (ILP) systems have been successfully applied to solve complex biological problem by viewing them as binary classification tasks. It remains an open question how an accurate solution to a multi-class problem can be obtained by using a logic based learning method. In this paper we present a novel logic based approach to solve complex and challenging multi-class classification problems in bioinformatics by focusing on a particular task, namely protein fold recognition. Our technique is based on the use of large margin kernel-based methods in conjunction with first order rules induced by an ILP system. The proposed approach learns a multi-class classifier by using a divide and conquer reduction strategy that splits multi-classes into binary groups and solves each individual problem recursively hence generating an underlying decision list structure. The method is applied to assigning protein domains to folds. Experimental evaluation of the method demonstrates the efficacy of the proposed approach to solving complex multi-class classification problems in bioinformatics. © 2009 ACM.
Kelley LA, Shrimpton PJ, Muggleton SH, et al., 2009, Discovering rules for protein-ligand specificity using support vector inductive logic programming, PROTEIN ENGINEERING DESIGN & SELECTION, Vol: 22, Pages: 561-567, ISSN: 1741-0126
- Author Web Link
- Cite
- Citations: 7
Tamaddoni-Nezhad A, Muggleton S, 2009, The lattice structure and refinement operators for the hypothesis space bounded by a bottom clause, MACHINE LEARNING, Vol: 76, Pages: 37-72, ISSN: 0885-6125
- Author Web Link
- Open Access Link
- Cite
- Citations: 13
Santos JCA, Tamaddoni-Nezhad A, Muggleton S, 2009, An ILP System for Learning Head Output Connected Predicates, 14th Portuguese Conference on Artificial Intelligence, Publisher: SPRINGER-VERLAG BERLIN, Pages: 150-159, ISSN: 0302-9743
- Author Web Link
- Cite
- Citations: 3
Lodhi H, Muggleton S, Sternberg MJE, 2009, Learning Large Margin First Order Decision Lists for Multi-Class Classification, 12th International Conference on Discovery Science, Publisher: SPRINGER-VERLAG BERLIN, Pages: 168-+, ISSN: 0302-9743
- Author Web Link
- Cite
- Citations: 3
Dietterich TG, Domingos P, Getoor L, et al., 2008, Structured machine learning: the next ten years, MACHINE LEARNING, Vol: 73, Pages: 3-23, ISSN: 0885-6125
- Author Web Link
- Cite
- Citations: 70
Chen J, Muggleton S, Santos J, 2008, Learning probabilistic logic models from probabilistic examples, MACHINE LEARNING, Vol: 73, Pages: 55-85, ISSN: 0885-6125
- Author Web Link
- Open Access Link
- Cite
- Citations: 8
Tsunoyama K, Amini A, Sternberg MJE, et al., 2008, Scaffold hopping in drug discovery using inductive logic programming, JOURNAL OF CHEMICAL INFORMATION AND MODELING, Vol: 48, Pages: 949-957, ISSN: 1549-9596
- Author Web Link
- Cite
- Citations: 30
Chen J, Kelley L, Muggleton S, et al., 2008, Protein fold discovery using stochastic logic programs, Pages: 244-262, ISSN: 0302-9743
This chapter starts with a general introduction to protein folding. We then present a probabilistic method of dealing with multi-class classification, in particular multi-class protein fold prediction, using Stochastic Logic Programs (SLPs). Multi-class prediction attempts to classify an observed datum or example into its proper classification given that it has been tested to have multiple predictions. We apply an SLP parameter estimation algorithm to a previous study in the protein fold prediction area, in which logic programs have been learned by Inductive Logic Programming (ILP) and a large number of multiple predictions have been detected. On the basis of several experiments, we demonstrate that PILP approaches (eg. SLPs) have advantages for solving multi-class (protein fold) prediction problems with the help of learned probabilities. In addition, we show that SLPs outperform ILP plus majority class predictor in both predictive accuracy and result interpretability. © 2008 Springer-Verlag Berlin Heidelberg.
Muggleton S, Chen J, 2008, A behavioral comparison of some probabilistic logic models, Pages: 305-324, ISSN: 0302-9743
Probabilistic Logic Models (PLMs) are efficient frameworks that combine the expressive power of first-order logic as knowledge representation and the capability to model uncertainty with probabilities. Stochastic Logic Programs (SLPs) and Statistical Relational Models (SRMs), which are considered as domain frequency approaches, and on the other hand Bayesian Logic Programs (BLPs) and Probabilistic Relational Models (PRMs) (possible worlds approaches), are promising PLMs in the categories. This paper is aimed at comparing the relative expressive power of these frameworks and developing translations between them based on a behavioral comparison of their semantics and probability computation. We identify that SLPs augmented with combining functions (namely extended SLPs) and BLPs can encode equivalent probability distributions, and we show how BLPs can define the same semantics as complete, range-restricted SLPs. We further demonstrate that BLPs (resp. SLPs) can encode the relational semantics of PRMs (resp. SRMs). Whenever applicable, we provide inter-translation algorithms, present their soundness and give worked examples. © 2008 Springer-Verlag Berlin Heidelberg.
Chen J, Muggleton S, Santos J, 2008, Learning probabilistic logic models from probabilistic examples, Pages: 22-23, ISSN: 0302-9743
This paper describes research in Probabilistic Inductive Logic Programing (PILP). The question investigated is whether PILP should always be used to learn from categorical examples. The data sets used by most PILP systems and applications have non-probabilistic class values, like those used in ILP systems. The main reason for this is the lack of an obvious source of probabilistic class values. In this context, we investigate the use of Abductive Stochastic Logic Programs (SLPs) for metabolic network learning. One of the machine learning approaches, which has been used to model the inhibitory effect of various toxins in the metabolic network of rats, is abductive ILP [3]. A group of rats are injected with a toxin and the changes on the concentrations of a number of chemical compounds are monitored over time. The binary information on up/down regulations of metabolite concentrations is combined with background knowledge representing a subset of the KEGG metabolic diagrams. An abductive ILP program is used to suggest the inhibitory effects occurring in the network.c © 2008 Springer-Verlag Berlin Heidelberg.
Bang J-W, Crockford DJ, Holmes E, et al., 2008, Integrative top-down system metabolic modeling in experimental disease states via data-driven bayesian methods (vol 7, pg 497, 2008), JOURNAL OF PROTEOME RESEARCH, Vol: 7, Pages: 1352-1352, ISSN: 1535-3893
Muggleton S, Tamaddoni-Nezhad A, 2008, QG/GA: a stochastic search for Progol, 16th International Conference of Inductive Logic Programming, Publisher: SPRINGER, Pages: 121-133, ISSN: 0885-6125
- Author Web Link
- Cite
- Citations: 18
Muggleton S, Otero R, Colton S, 2008, Guest editorial: Special issue on inductive logic programming, MACHINE LEARNING, Vol: 70, Pages: 119-120, ISSN: 0885-6125
Bang J-W, Crockford DJ, Hohmes E, et al., 2008, Integrative top-down system metabolic modeling in experimental disease states via data-driven Bayesian methods, JOURNAL OF PROTEOME RESEARCH, Vol: 7, Pages: 497-503, ISSN: 1535-3893
- Author Web Link
- Cite
- Citations: 15
Bang J-W, Crockford DJ, Holmes E, et al., 2008, Integrative top-down system metabolic modeling in experimental disease states via data-driven Bayesian methods., J Proteome Res, Vol: 7, Pages: 497-503, ISSN: 1535-3893
Multivariate metabolic profiles from biofluids such as urine and plasma are highly indicative of the biological fitness of complex organisms and can be captured analytically in order to derive top-down systems biology models. The application of currently available modeling approaches to human and animal metabolic pathway modeling is problematic because of multicompartmental cellular and tissue exchange of metabolites operating on many time scales. Hence, novel approaches are needed to analyze metabolic data obtained using minimally invasive sampling methods in order to reconstruct the patho-physiological modulations of metabolic interactions that are representative of whole system dynamics. Here, we show that spectroscopically derived metabolic data in experimental liver injury studies (induced by hydrazine and alpha-napthylisothiocyanate treatment) can be used to derive insightful probabilistic graphical models of metabolite dependencies, which we refer to as metabolic interactome maps. Using these, system level mechanistic information on homeostasis can be inferred, and the degree of reversibility of induced lesions can be related to variations in the metabolic network patterns. This approach has wider application in assessment of system level dysfunction in animal or human studies from noninvasive measurements.
Muggleton S, 2008, From ILP to PILP, ISSN: 0302-9743
Inductive Logic Programming (ILP) is the area of Computer Science which deals with the induction of hypothesised predicate definitions from examples and background knowledge. Probabilistic ILP (PILP) extends the ILP framework by making use of probabilistic variants of logic programs to capture background and hypothesised knowledge. ILP and PILP are differentiated from most other forms of Machine Learning (ML) both by their use of an expressive representation language and their ability to make use of logically encoded background knowledge. This has allowed successful applications in areas such as Systems Biology, computational chemistry and Natural Language Processing. The problem of learning a set of logical clauses from examples and background knowledge has been studied since Reynold's and Plotkin's work in the late 1960's. The research area of ILP has been studied intensively since the early 1990s, while PILP has received increasing amounts of interest over the last decade. This talk will provide an overview of results for learning logic programs within the paradigms of learning-in-the-limit, PAC-learning and Bayesian learning. These results will be related to various settings, implementations and applications used in ILP. It will be argued that the Bayes' setting has a number of distinct advantages for both ILP and PILP. Bayes' average case results are easier to compare with empirical machine learning performance than results from either PAC or learning-in-thelimit. Broad classes of logic programs are learnable in polynomial time in a Bayes' setting, while corresponding PAC results are largely negative. Bayes' can be used to derive and analyse algorithms for learning from positive only examples for classes of logic program which are unlearnable within both the PAC and learning-in-the-limit framework. It will be shown how a Bayesian approach can be used to analyse the relevance of background knowledge when learning. General results will also be discussed for expec
Muggleton S, 2008, Developing robust synthetic Biology designs using a microfluidic robot scientist, Pages: 4-5, ISSN: 0302-9743
Synthetic Biology is an emerging discipline that is providing a conceptual framework for biological engineering based on principles of standardisation, modularity and abstraction. For this approach to achieve the ends of becoming a widely applicable engineering discipline it is critical that the resulting biological devices are capable of functioning according to a given specification in a robust fashion. In this talk we will describe the development of techniques for experimental validation and revision based on the development of a microfluidic robot scientist to support the empirical testing and automatic revision of robust component and device-level designs. The approach is based on probabilistic and logical hypotheses [1] generated by active machine learning. Previous papers [2,3] based on the author's design of a Robot Scientist appeared in Nature and was widely reported in the press. The new techniques will extend those in the speaker's previous publications in which it was demonstrated that the scientific cycle of hypothesis formation, choice of low-expected cost experiments and the conducting of biological experiments could be implemented in a fully automated closedloop. In the present work we are developing the use of Chemical Turing machines based on micro-fluidic technology, to allow high-speed (sub-second) turnaround in the cycle of hypothesis formation and testing. If successful such an approach should allow a speed-up of several orders of magnitude compared to the previous technique (previously 24 hour experimental cycle). © 2008 Springer Berlin Heidelberg.
Muggleton SH, 2008, From ILP to PILP., Publisher: Springer, Pages: 7-7
Muggleton SH, Santos JCA, Tamaddoni-Nezhad A, 2008, TopLog: ILP Using a Logic Program Declarative Bias, 24th International Conference on Logic Programming (ICLP), Publisher: SPRINGER-VERLAG BERLIN, Pages: 687-692, ISSN: 0302-9743
- Author Web Link
- Cite
- Citations: 18
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.